Chapter 7: Building Theories Fit For Design
Useful theory is notoriously difficult to build in HCI. This
chapter examines why this is so, suggests ways that
RODS,
HASTI, and
CoSTH can potentially solve these difficulties,
and explores ways of making these theoretical resources more usable
to researchers and practitioners seeking to apply them in real design
situations.
The main suggestion for why HCI theory has had so little impact is that
they produce the wrong sort of results: they concentrate on usability
and performance, particularly detailed aspects of these. The theories
and frameworks pursued here, in contrast, persue usefulness in a
very broad-brush and general manner. These types of resources are
most useful and cost-effective for designers trying to synthesize a new
design from an understanding of the design context. We need design
theories. Cognitive support theories can be converted into just
this type of theory. Prior HCI work has tended to explain or predict
problems when one has a design (or tool) in hand. As part of the argument
promoting cognitive support theories as design theories, an analysis is
presented of existing theories in HCI, and a taxonomy is introduced of
design knowledge based on an analysis of the cognitive processes involved
in design.
In addition, the question of evaluating design theories is covered.
Methods for evaluating design theory are poorly understood. Several
different evaluation criteria and methods are proposed. It is shown
that several of these evaluation methods are performed within other
chapters.
Finally, this chapter considers the issue of theory usability.
Besides being useful for designers, the theories must be usable.
Often this requirement means that a full-blown theory cannot be used
since it could be difficult to use and costly to apply. Instead,
lightweight and specialized resources must be produced. This chapter
explores three different ways of converting the cognitive support
theories into usable components of a designer's toolkit. These are:
(1) the generation of a suitable vocabulary and associated concepts,
(2) the provision of design stances that can be used by the designer to
generate design goals that they might not have otherwise thought of, and
(3) the construction of specialized checklists and tables that can be
used to methodologically consider design choices. These (or similar)
theoretically-inspired derived resources might some day be usefully
taught in tutorials and undergraduate courses.
|